The problem of how to restore mobile sensors' energy efficiently has become a very important field of investigation in Mobile Sensor Networks (MSN). The main goal of any energy restoration technique is to reach a state where the continuous operation of the sensor network can be guaranteed and the number of sensor losses can be limited in scope and time. The present thesis analyzes the impact of adding physical obstacles to the MSN scenarios. It is proposed that an efficient obstacle avoidance technique could be introduced that allows sensors to collect important information about the obstacles. This information could be used by the sensors to improve their paths and avoid unnecessary exploration. Different energy restoration techniques in presence of physical obstacles are compared. It is shown in the simulation results that our proactive strategy, which uses information about the obstacles to improving the paths around them, achieves the best results.